PURPOSE (2024)

The Applied Digital Islamicate Studies Educational Program aims to contribute to the digitalizing of Islamicate research's processes. In this context, it was initially organized in February 2024, with the partnership's CDH in Marmara University and supported by TUBITAK, as an educational program for researchers in Ilahiyat faculties* (in Turkey), focusing on the fundamental methods of digital humanities. Modules covering geographical information systems, network analyses, text analyses, and artificial intelligence were incorporated into the program, bringing together researchers pursuing master's and doctoral degrees in those faculties and expert educators in the field. As a result, participant preferences in the artificial intelligence (AI) module during this training led to the necessity of focusing the second training program on developments in the use of AI methods in Islamicate research. The target audience consists of researchers pursuing master's and doctoral degrees in Ilahiyat faculties, and a new 7-day training program has been organized between 22 July and 5 August.

Artificial Intelligence (AI) has long been a captivating field in the scientific community, representing an expansive domain with numerous subfields. AI methods find applications in various areas, ranging from medical research to environmental engineering, and from autonomous driving technologies to facial recognition systems. Within the realm of education, AI has established its presence through applications such as translation services, summarization tools, and voice assistants, manifested in platforms providing translation services, digital dictionaries, language learning applications, and AI design assistants. The logic of human intelligence, involving memorization, repetition, and pattern recognition, is being digitally reprocessed through AI. The success of machines in emulating human learning processes is progressively gaining momentum, and the profound impact of the AI field, with its extensive history, is notably evident in the social sciences. Artificial neural networks, natural language processing, optical character recognition, and machine learning algorithms are increasingly integrated into data-driven research processes within the social sciences. The emulation of human language by machines, the automatic transcription of handwritten texts, and the application of research studies to texts for data acquisition and analysis represent novel horizons opening up for researchers.

The rapid transformation in digitalization appears to necessitate the inevitable adaptation of theological research. Studies within the Islamicate studies utilizing methods from artificial intelligence subdomains such as machine learning, text mining, deep learning, and natural language processing have gained momentum in recent years. The utilization of Natural Language Processing (NLP) technologies in research focused on the Arabic language, as seen in projects like CAMel-Lab, enables the digital transfer of the linguistic wealth of the broader Arab region. Platforms based on AI methods, including OCR and HTR technologies such as Transkribus and Akis, have shown significant progress, particularly in examining Ottoman Turkish texts. Platforms like eScriptorium, which center their work on manuscript studies, have employed OpenITI models, aiming to enrich Islamic literature with new texts (Saraçoğlu, 2023, 863). Beyond the usage of digital dictionaries for Ottoman Turkish, Lexiqamus stands out as a service based on OCR technologies, where words are transcribed in Arabic script, offering a tool for researchers focusing on manuscript works. Additionally, "ukhBERT," an NLP model developed by Mohamed Alkaoud and Mairaj Syed, specifically addresses the identification of narrators' names in the chains of transmission (isnād). This study aims to resolve challenges arising from the multiple spellings of narrator names in the isnād due to the structure of Arabic names, leading to ambiguities in information transmission (Alkaoud & Syed, 2010).

The utilization of AI methods in theological research, requiring a robust technical infrastructure and substantial financial support, is still limited. Considering the vast wealth of Islamic literature that could form an extensive data repository, this field holds significant potential for making substantial contributions to the representation of Islamic literature both nationally and internationally. Some of the limited publications and projects in the field are outlined below:

Publications:

Aqil M. Azmi and Nawaf bin Bedia, "e-Narrator-An Application for Creating an Ontology of Hadiths Narration Tree Semantically and Graphically," The Arabian Journal for Science and Engineering, 35/2C (2010), 51-68;
Aqil M. Azmi and Amjad M. AlOfaidly, "A Novel Method to Automatically Pass Hukm on Hadith," 5th International Conference on Arabic Language Processing (CITALA '14), Fes 2014.
Nizar Y. Habash, Introduction to Arabic Natural Language Processing, thk. Graeme Hirst (Morgan & Claypool, 2010); Imed Zitouni (ed.), Natural Language Processing of Semitic Languages, thk. Graeme Hirst vd. (Berlin:Springer, 2014).
Tombul, Sema. “Veri Madenciliği Tekniklerinin ve Algoritmik Araştırmaların Hadis İlmine Uygulanabilirliği”, Eskiyeni, 44 (2021), 461-474.
Mohamed Alkaoud- Mairaj Syed, “Learning to Identify Narrators in Classical Arabic Texts”, Procedia Computer Science, 189 (2021), 335-342.

Projects:

“"Utilization of Artificial Intelligence Methods in the Comparative Analysis of Ottoman Period Fatwa Compilations," Supported by TÜBİTAK, Principal Investigator: Dr. Res. Assist. Ahmet Faruk Çelik (ongoing). As of June 2022, this TÜBİTAK-supported project is based on text analyses. The project aims to develop an artificial intelligence program that allows the comparison of fatwas from Minkârizâde Yahya Efendi, a 17th-century Sheikh ul-Islam, with fatwas of other Sheikh ul-Islams. Through these comparisons, the project seeks to question the similarity rates of fatwa compilations that, at first glance, exhibit structural similarities, exploring commonalities based on their structural elements(which will be covered in this training).
“Linked Open Tafsir" is a project led by Ömer Özsoy from Goethe University's Academy for Islam in Science and Society (AIWG) in Frankfurt, Serdar Kurnaz from Humboldt University in Berlin, and Yaşar Sarıkaya from Justus Liebig University in Giessen. The project, conducted in collaboration with these three universities between 2018 and 2022, focuses on coding and categorizing the narrations and chains of transmission of Ibn Jarir al-Tabari's (d. 310/923) work "Jamiu'l-Bayan an Te'wil Ayyi'l-Qur'an." After employing text analysis methods in this stage, the obtained data is reprocessed and presented to researchers using artificial intelligence algorithms to address specific research questions. In this regard, the project stands out as a unique structure, not only utilizing artificial intelligence algorithms in the data collection phase but also incorporating them as a framework for further analysis.
“Quan in Fiqh," led by Yusuf Çelik and Christian Lange, is an intriguing project in the field of theology, known as Footprinter (Qur’an in Fıqh). This project stands out for its ease of use, graphical readability, visualization preferences, and direct access to data in studies based on text analysis in theology. In this project, using text analysis, verses in fiqh texts and the relationships between verses and sections in texts by scholars from different schools of thought are visualized. The main texts are classified according to schools of thought to facilitate the researcher's ability to read the verses within the context. This allows for the efficient identification of extensive data that would otherwise take a long time for a commentary researcher. The website enables queries based on verse number, author name, text name, historical period, and schools of thought (which will be covered in this training).

In the past few years, the emergence of large language models has brought new perspectives to the inclusion of digital research methods in the education processes of theology faculties. The development of the HadithGPT model, based on texts from the field of hadith within a few months of OpenAI releasing ChatGPT, is noteworthy. Similarly, the creation of the IsarGPT model in Turkey in 2023 (which will be covered in this training) indicates the rapid impact of large language models on theological research.

The mentioned developments, on the one hand, demonstrate the potential integration of Artificial Intelligence methods into theological studies, but on the other hand, they also bring about a significant adaptation challenge. Within the scope of this project, the aim is to create a learning environment for researchers in Islamicate studies, address methodological problems from both Artificial Intelligence and theological perspectives using examples, and introduce contemporary and rapidly evolving methods to the agenda of Islamicate researchers.

The project aims to provide participants with an interdisciplinary and multicultural collaborative environment where educators from various universities in Turkey and around the world are involved. The broad dissemination of the training to a diverse audience is expected to contribute to the emergence of collaborative projects, the formation of a network among researchers interested in the field, and the fostering of joint efforts. Encouraging participants to directly publish their individual or group works during or after the project on the "Project Outputs" section of the project website or providing links to their relevant publications on the site is a common goal. Additionally, in the event of project approval, priority will be given to individuals who participated in the training in previous periods.

* In Turkey, Ilahiyat faculties contribute to the body of knowledge in disciplines like Islamic studies, history of Islam, sociology of religion, psychology of religion, Islamic arts, and Islamic philosophy.